Blame research design for failed HIV study

In the past decade, the medical community has made great strides in HIV prevention methods. It’s a promising landscape, with studies showing the effectiveness of anti-retrovirals (ARV) as a base for pills, topical gels and barrier methods that men and women can use to protect themselves from HIV.

Unfortunately, a study published recently in the prestigious New England Journal of Medicine (NEJM) represents a step back. Known as VOICE, or Vaginal and Oral Interventions to Control the Epidemic, the study failed to show any preventive results for women in southern Africa using ARV-based pre-exposure prophylaxis (PrEP) pills or topical microbicide gel.

It’s a particularly unsettling failure because previous studies have demonstrated that these ARV-based methods work. Most of the women who participated in the VOICE study did not use the tablets or gel, but those who did were protected. In other words, the study failed not because the products didn’t work but because they weren’t used.

The scientific community has suspected the VOICE study would present disappointing results since 2011, when the research — involving more than 5,000 women and costs of $94 million — was discontinued for futility. The NEJM authors’ main explanation for the failure is the participants’ deceit and lack of compliance.

But when a study fails, we must be careful not to imply that the subjects are at fault. My analysis of the study suggests, rather, that research design was to blame. The challenge of this research is more social and behavioral than medical; to succeed, we must better understand which routines and methods work best for women in stressful daily conditions. If the offered methods are not used, then researchers must rethink their approach or at-risk women will continue to become infected with HIV, and the epidemic will spiral. 

In any clinical trial, it is the responsibility of researchers to build trust. Without long-term preparation, any community will be suspicious of novel interventions. Especially in South Africa, where many of the VOICE participants were recruited, there are steep cultural challenges that researchers must confront, including more than 40 years of apartheid and a decade of AIDS denialism supported by the Department of Health.

In spite of this history, a grass-roots and globally supported treatment action campaign, which included the efforts of Nelson Mandela in his last years, led to a change in government policy and the provision of ARV treatment by the state health system. When the program was rolled out, millions of South Africans adhered consistently to treatment. Many lives have been saved and others protected from infection.

Any new research requires community involvement and rigorous assessment of practices. The VOICE researchers could have learned from a successful microbicide trial, CAPRISA 004, run by the Center for the AIDS Program of Research in South Africa. After conducting interviews, researchers decided to ask women to use the microbicide gel only before and after they had sex, rather than daily. This took into account that many of their partners were migrant workers who were away for long periods and that women might not stick to an unnecessary daily routine, knowing they were unlikely to have sex. The researchers established an interactive feedback system: Women were asked to return the microbicide gel applicators (which were tested as the study proceeded to detect whether they had been appropriately used) and to meet with motivational counselors. In 2010 the CAPRISA trial demonstrated that many women used the microbicide on the day they had sex and that it was protective.

Rather than follow this precedent, the VOICE researchers asked its participants to use the microbicide daily. The trial failed because the women did not use the products as directed. The NEJM article focuses on the fact that women did not always disclose that they had not taken the pills or used the gel and handed in pill bottles and applicators that appeared to have been used — even though later blood testing showed no traces of the study drug. With more effective ongoing biomarkers and feedback mechanisms, these problems might have surfaced and been remedied earlier.

The VOICE study appeared to lack sufficient community involvement. CAPRISA 004 focused on women in Durban (as did VOICE) and in Vulindlela, a rural area outside Durban. First securing the support of local Zulu chiefs, CAPRISA developed widespread community engagements in Vulindlela. Doctors attended local council events, and the clinic provided AIDS treatment. It is significant that CAPRISA 004’s results were much better in Vulindlela than they were in Durban. Women in rural areas have historically been more constricted by patriarchal relations, more suspicious of drugs dubbed toxic by traditional leaders and, therefore, less likely to adhere to preventive measures than urban-based women. The data suggest that the long-term, substantive community engagement in Vulindlela boosted women’s cooperation. Rather than build on these important successes, VOICE apparently did not adequately consider the social situations of the women in its study.

Bottom-up research design may improve results, but this takes time, costs money and disrupts accepted hierarchies. Because funders and donors may not recognize the need to build in the costs of community engagement, studies are more likely to focus on pharmaceutical methods than on strong investment in local participation. But as we have learned from decades of AIDS research and recently with Ebola, without ongoing community involvement, it is not possible to effectively address an epidemic.

The high HIV incidence observed in the VOICE study remind us that we desperately need methods women can more easily put into practice in difficult situations, especially those most at risk in southern Africa: women ages 15 to 22. There are at least two promising products on the horizon — one a monthly ring, the other an injection administered four times a year — that could become additional, vital tools in fighting this ongoing crisis.

Meanwhile, PrEP is no panacea. Not only are there side effects, but also people need to use it appropriately. But making PrEP available everywhere women are at risk — not only to women in the U.S. — is nonetheless a crucial next step. Most important, to reduce the incidence of HIV infection globally, we must work with women and men at the community level to promote trust and identify the best ways to use the effective methods we already have.